Detection of player learning curve in a car driving game

Boyan Bontchev, Dessislava Vassileva

    Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingAcademicpeer-review

    Abstract

    A learning curve (called also experience curve, or productivity curve) provides a graphical representation of progress in learning over time. In video games, the learning curve illustrates how players must spend time for mastering the game by both developing playing skills and acquiring cognitive abilities and knowledge for solving the game challenges. Naturally, learning curves vary from game to game and, as well, from player to player. A player learning curve represents the progress in learning and, hence, performance specific to an individual player. Therefore, detection of the player learning curve appears to be crucial for the realization of an effective gameplay because it reflects player motivation and engagement and can be applied for further adjusting of game difficulty and challenges.

    The paper presents a way of automatic detection of player learning curves as behavioral patterns of player performance, by applying a software component called “Player-centric rule-and-pattern-based adaptation asset” developed in the scope of the RAGE (Realising and Applied Gaming Ecosystem) H2020 project. The component is integrated within a car driving video game in order to detect dynamically specific player learning curves representing overall player performance (OPP) over time. The game was developed by the authors of the paper and provides a fascinating first-person car driving simulation with maneuvering on a road with various types of curves, through attracting and realistic terrains. It increases player motivation and fun by visualization of extra tools like a velometer and trip counter, by playing amazing sounds (of the motor, of each collision, etc.), and by driving under different weather conditions such as fog, rain, and at nightfall. The game registers the player metric of OPP calculated as:
    OPP = Vav / (Iav_norm + 1),
    where Vav is the average velocity for a time window of 10 seconds, and Iav_norm is the average normalized impact of collisions:
    Iav_norm = SUM(V * |sin(α)|) / Ncoll / Iav,
    where V is the car velocity at the moment of collision, α is the angle between the car direction and the hit body, Ncoll is the number of collisions for the time window, and Iav is the average impact value for the game. The software component allows registration of patterns of OPP by simple formal definitions including the setting of:
    • Key points of the value of chosen player metric at absolute or relative moments of time
    • Pattern search accuracy – meaning % of difference from the pattern points
    • Pattern search recall – meaning % of points being out of the pattern area
    In case of a found occurrence of specific pattern for the metric or for its feature (such as mean, deviation, or moving average), the component executes an event handler, which can be customised by the game developer in order to realise a specific player control or adaptation of some game metrics according to that change. The paper shows how the component was applied for detection of learning curve types possible for the car driving game. The curves present the evolution of OPP and its mean and moving average values when applying various driving conditions such as fog, rain, and nightfall. The detected player learning curve will be further used for adjusting of the game mechanics (friction to the road at the rain, or/and visibility at the fog, or/and illumination at nightfall) for the individual player, for providing better fun and immersion of the game.
    Original languageEnglish
    Title of host publication12th International Technology, Education and Development Conference Proceedings
    Subtitle of host publicationINTED2018 Proceedings
    PublisherIATED Academy
    ISBN (Print) 978-84-697-9480-7
    DOIs
    Publication statusPublished - 2018
    Event 12th International Technology, Education and Development Conference: INTED2018 - Valencia, Spain
    Duration: 5 Mar 20187 Mar 2018
    https://library.iated.org/publications/INTED2018

    Conference

    Conference 12th International Technology, Education and Development Conference
    CountrySpain
    CityValencia
    Period5/03/187/03/18
    Internet address

    Fingerprint

    Railroad cars
    Fog
    Rain
    Visibility
    Ecosystems
    Mechanics
    Visualization
    Lighting
    Productivity
    Acoustic waves
    Friction

    Keywords

    • player learning curve
    • detection
    • pattern
    • component
    • RAGE

    Cite this

    Bontchev, B., & Vassileva, D. (2018). Detection of player learning curve in a car driving game. In 12th International Technology, Education and Development Conference Proceedings: INTED2018 Proceedings IATED Academy. https://doi.org/10.21125/inted.2018.2288
    Bontchev, Boyan ; Vassileva, Dessislava. / Detection of player learning curve in a car driving game. 12th International Technology, Education and Development Conference Proceedings: INTED2018 Proceedings. IATED Academy, 2018.
    @inproceedings{3876be3f628140b9a047e14f0dd07efb,
    title = "Detection of player learning curve in a car driving game",
    abstract = "A learning curve (called also experience curve, or productivity curve) provides a graphical representation of progress in learning over time. In video games, the learning curve illustrates how players must spend time for mastering the game by both developing playing skills and acquiring cognitive abilities and knowledge for solving the game challenges. Naturally, learning curves vary from game to game and, as well, from player to player. A player learning curve represents the progress in learning and, hence, performance specific to an individual player. Therefore, detection of the player learning curve appears to be crucial for the realization of an effective gameplay because it reflects player motivation and engagement and can be applied for further adjusting of game difficulty and challenges.The paper presents a way of automatic detection of player learning curves as behavioral patterns of player performance, by applying a software component called “Player-centric rule-and-pattern-based adaptation asset” developed in the scope of the RAGE (Realising and Applied Gaming Ecosystem) H2020 project. The component is integrated within a car driving video game in order to detect dynamically specific player learning curves representing overall player performance (OPP) over time. The game was developed by the authors of the paper and provides a fascinating first-person car driving simulation with maneuvering on a road with various types of curves, through attracting and realistic terrains. It increases player motivation and fun by visualization of extra tools like a velometer and trip counter, by playing amazing sounds (of the motor, of each collision, etc.), and by driving under different weather conditions such as fog, rain, and at nightfall. The game registers the player metric of OPP calculated as: OPP = Vav / (Iav_norm + 1), where Vav is the average velocity for a time window of 10 seconds, and Iav_norm is the average normalized impact of collisions: Iav_norm = SUM(V * |sin(α)|) / Ncoll / Iav, where V is the car velocity at the moment of collision, α is the angle between the car direction and the hit body, Ncoll is the number of collisions for the time window, and Iav is the average impact value for the game. The software component allows registration of patterns of OPP by simple formal definitions including the setting of:• Key points of the value of chosen player metric at absolute or relative moments of time• Pattern search accuracy – meaning {\%} of difference from the pattern points• Pattern search recall – meaning {\%} of points being out of the pattern areaIn case of a found occurrence of specific pattern for the metric or for its feature (such as mean, deviation, or moving average), the component executes an event handler, which can be customised by the game developer in order to realise a specific player control or adaptation of some game metrics according to that change. The paper shows how the component was applied for detection of learning curve types possible for the car driving game. The curves present the evolution of OPP and its mean and moving average values when applying various driving conditions such as fog, rain, and nightfall. The detected player learning curve will be further used for adjusting of the game mechanics (friction to the road at the rain, or/and visibility at the fog, or/and illumination at nightfall) for the individual player, for providing better fun and immersion of the game.",
    keywords = "player learning curve, detection, pattern, component, RAGE",
    author = "Boyan Bontchev and Dessislava Vassileva",
    year = "2018",
    doi = "10.21125/inted.2018.2288",
    language = "English",
    isbn = "978-84-697-9480-7",
    booktitle = "12th International Technology, Education and Development Conference Proceedings",
    publisher = "IATED Academy",

    }

    Bontchev, B & Vassileva, D 2018, Detection of player learning curve in a car driving game. in 12th International Technology, Education and Development Conference Proceedings: INTED2018 Proceedings. IATED Academy, 12th International Technology, Education and Development Conference, Valencia, Spain, 5/03/18. https://doi.org/10.21125/inted.2018.2288

    Detection of player learning curve in a car driving game. / Bontchev, Boyan; Vassileva, Dessislava.

    12th International Technology, Education and Development Conference Proceedings: INTED2018 Proceedings. IATED Academy, 2018.

    Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingAcademicpeer-review

    TY - GEN

    T1 - Detection of player learning curve in a car driving game

    AU - Bontchev, Boyan

    AU - Vassileva, Dessislava

    PY - 2018

    Y1 - 2018

    N2 - A learning curve (called also experience curve, or productivity curve) provides a graphical representation of progress in learning over time. In video games, the learning curve illustrates how players must spend time for mastering the game by both developing playing skills and acquiring cognitive abilities and knowledge for solving the game challenges. Naturally, learning curves vary from game to game and, as well, from player to player. A player learning curve represents the progress in learning and, hence, performance specific to an individual player. Therefore, detection of the player learning curve appears to be crucial for the realization of an effective gameplay because it reflects player motivation and engagement and can be applied for further adjusting of game difficulty and challenges.The paper presents a way of automatic detection of player learning curves as behavioral patterns of player performance, by applying a software component called “Player-centric rule-and-pattern-based adaptation asset” developed in the scope of the RAGE (Realising and Applied Gaming Ecosystem) H2020 project. The component is integrated within a car driving video game in order to detect dynamically specific player learning curves representing overall player performance (OPP) over time. The game was developed by the authors of the paper and provides a fascinating first-person car driving simulation with maneuvering on a road with various types of curves, through attracting and realistic terrains. It increases player motivation and fun by visualization of extra tools like a velometer and trip counter, by playing amazing sounds (of the motor, of each collision, etc.), and by driving under different weather conditions such as fog, rain, and at nightfall. The game registers the player metric of OPP calculated as: OPP = Vav / (Iav_norm + 1), where Vav is the average velocity for a time window of 10 seconds, and Iav_norm is the average normalized impact of collisions: Iav_norm = SUM(V * |sin(α)|) / Ncoll / Iav, where V is the car velocity at the moment of collision, α is the angle between the car direction and the hit body, Ncoll is the number of collisions for the time window, and Iav is the average impact value for the game. The software component allows registration of patterns of OPP by simple formal definitions including the setting of:• Key points of the value of chosen player metric at absolute or relative moments of time• Pattern search accuracy – meaning % of difference from the pattern points• Pattern search recall – meaning % of points being out of the pattern areaIn case of a found occurrence of specific pattern for the metric or for its feature (such as mean, deviation, or moving average), the component executes an event handler, which can be customised by the game developer in order to realise a specific player control or adaptation of some game metrics according to that change. The paper shows how the component was applied for detection of learning curve types possible for the car driving game. The curves present the evolution of OPP and its mean and moving average values when applying various driving conditions such as fog, rain, and nightfall. The detected player learning curve will be further used for adjusting of the game mechanics (friction to the road at the rain, or/and visibility at the fog, or/and illumination at nightfall) for the individual player, for providing better fun and immersion of the game.

    AB - A learning curve (called also experience curve, or productivity curve) provides a graphical representation of progress in learning over time. In video games, the learning curve illustrates how players must spend time for mastering the game by both developing playing skills and acquiring cognitive abilities and knowledge for solving the game challenges. Naturally, learning curves vary from game to game and, as well, from player to player. A player learning curve represents the progress in learning and, hence, performance specific to an individual player. Therefore, detection of the player learning curve appears to be crucial for the realization of an effective gameplay because it reflects player motivation and engagement and can be applied for further adjusting of game difficulty and challenges.The paper presents a way of automatic detection of player learning curves as behavioral patterns of player performance, by applying a software component called “Player-centric rule-and-pattern-based adaptation asset” developed in the scope of the RAGE (Realising and Applied Gaming Ecosystem) H2020 project. The component is integrated within a car driving video game in order to detect dynamically specific player learning curves representing overall player performance (OPP) over time. The game was developed by the authors of the paper and provides a fascinating first-person car driving simulation with maneuvering on a road with various types of curves, through attracting and realistic terrains. It increases player motivation and fun by visualization of extra tools like a velometer and trip counter, by playing amazing sounds (of the motor, of each collision, etc.), and by driving under different weather conditions such as fog, rain, and at nightfall. The game registers the player metric of OPP calculated as: OPP = Vav / (Iav_norm + 1), where Vav is the average velocity for a time window of 10 seconds, and Iav_norm is the average normalized impact of collisions: Iav_norm = SUM(V * |sin(α)|) / Ncoll / Iav, where V is the car velocity at the moment of collision, α is the angle between the car direction and the hit body, Ncoll is the number of collisions for the time window, and Iav is the average impact value for the game. The software component allows registration of patterns of OPP by simple formal definitions including the setting of:• Key points of the value of chosen player metric at absolute or relative moments of time• Pattern search accuracy – meaning % of difference from the pattern points• Pattern search recall – meaning % of points being out of the pattern areaIn case of a found occurrence of specific pattern for the metric or for its feature (such as mean, deviation, or moving average), the component executes an event handler, which can be customised by the game developer in order to realise a specific player control or adaptation of some game metrics according to that change. The paper shows how the component was applied for detection of learning curve types possible for the car driving game. The curves present the evolution of OPP and its mean and moving average values when applying various driving conditions such as fog, rain, and nightfall. The detected player learning curve will be further used for adjusting of the game mechanics (friction to the road at the rain, or/and visibility at the fog, or/and illumination at nightfall) for the individual player, for providing better fun and immersion of the game.

    KW - player learning curve

    KW - detection

    KW - pattern

    KW - component

    KW - RAGE

    U2 - 10.21125/inted.2018.2288

    DO - 10.21125/inted.2018.2288

    M3 - Conference article in proceeding

    SN - 978-84-697-9480-7

    BT - 12th International Technology, Education and Development Conference Proceedings

    PB - IATED Academy

    ER -

    Bontchev B, Vassileva D. Detection of player learning curve in a car driving game. In 12th International Technology, Education and Development Conference Proceedings: INTED2018 Proceedings. IATED Academy. 2018 https://doi.org/10.21125/inted.2018.2288